Minoh Jeong is a postdoctoral researcher whose work sits at the intersection of machine learning, statistical signal processing, and information theory. He designs principled algorithms that turn raw, high-dimensional data—images, sensor streams, genomic profiles—into compact, meaningful representations. A central thread of his research is contrastive learning, an approach that “teaches by comparison.” By pulling semantically related examples closer together and pushing unrelated ones apart, his methods uncover the hidden structure of complex data while remaining robust to noise and missing labels.
His work bridges rigorous mathematical theory with practical impact, offering new tools for tasks such as noisy-label learning, probabilistic representation, and multimodal alignment. By participating in MIDAS, he seeks to collaborate across disciplines where data complexity and interpretability are essential, contributing methodological advances to areas such as bioinformatics, vision-language modeling, and beyond.